| Can a resonance create issues even if it stays below the required impedance budget? |
Yes, a resonance can still cause problems even if it stays below the required impedance budget.
Here’s why: you might see a relatively flat impedance profile, but if a resonance peak shows up, even below the budget line, it can introduce issues such as poor regulation or stability problems. The most common risk comes from a high-Q resonance—because even though it doesn’t technically exceed the target impedance, it can still interact with the system in a way that disrupts performance.
It’s definitely something to watch for. Tools like NISM (Non-Invasive Stability Measurement) are very useful here because they let you quickly analyze the impedance-versus-frequency trace and determine whether a given resonance peak is likely to cause instability. |
| What are the best methods to reduce EMI in high-power SMPS designs? |
One of the most effective ways to reduce EMI in high-power SMPS designs is to manage ground currents carefully. Preventing common-mode currents is especially important, and this comes down to solid ground plane design.
Another key factor is maintaining a low and flat PDN impedance profile. Resonances or areas where the impedance spikes can become sources of EMI, so controlling them is critical.
The challenge with high-power switch-mode power supplies is that they switch large amounts of current very quickly, essentially a recipe for EMI. By ensuring robust ground paths and keeping the impedance low and well-controlled across frequencies, you can significantly minimize EMI problems. |
| A few questions for Mike: 1) How important is impedance reduction and power integrity in power electronics design for high-power applications? 2) How did you start in the testing and measurement industry? 3) And what are the biggest challenges of maintaining signal integrity? |
It’s extremely important. The goal isn’t just to keep impedance low, but also to keep it flat across frequencies. A flatter impedance profile leads to better stability and reduced EMI, which are both critical in high-power applications. |
| How do you calculate the target impedance for a 400A load, current 2% ripple voltage for 1V? |
I would typically rely on software tools. They’re designed to handle the math and ensure accuracy, so using a dedicated tool is the most practical way to determine the target impedance for those conditions. |
| Is there a way to measure this noise for high-voltage systems? Like grid-level voltage, maybe 1kV? |
Yes, it is possible, but you need to proceed carefully. At grid-level voltages, such as 1 kV, you’d typically use high-attenuation probes on the order of 1000:1 or higher. These probes are designed with high dielectric strength, so you can measure safely without being too close to the voltage.
In high-voltage systems, ripple voltages are usually in the range of volts rather than millivolts, so they’re easier to measure with the right setup. Another option is to use isolated oscilloscopes, which don’t have a direct ground reference. This allows you to connect safely to high-voltage signals without risking user safety or damaging the equipment. |
| What is the best way to optimize transient response in a regulator that does not have compensation capability? |
For regulators that don’t have built-in compensation capability, optimizing transient response really comes down to managing impedance. One approach is to use fast-switching electronic loads to create transients that substitute for the device under test. By analyzing the step response, including any ringing, you can gain insight into the PDN impedance and identify potential stability issues.
The key is to minimize PDN impedance at the critical frequencies where transients occur. Doing so will significantly improve transient response even without direct compensation in the regulator. Tools such as RNS (Regulator Non-Invasive Stability) can also be useful for evaluating PDN impedance and stability. |
| Do you extend your analysis techniques up and beyond 7 GHz for wireless / RFI tolerance, near field, vs just EMI far field? |
Yes, we do extend our analysis well beyond 7 GHz. Using near-field probes, we can measure up to 40 GHz and beyond when needed. These types of measurements, however, are generally not applied to the power distribution network (PDN), since PDN analysis typically focuses on much lower frequencies. At higher frequencies, the measurements are more about wireless, RFI tolerance, and other signal-related effects.
We also perform near-field probing as part of this process. One point worth noting: when measuring PDN impedance using a two-port shunt method, that technique is only valid when the impedance is very low, on the order of a few ohms or less. Once the impedance rises above that, the method is no longer accurate. |
| Looks like R&S has a good probe for PDN impedance. Is there also a good/optimized emission sniffer probe that can be used with the same instruments (scope/VNA/SA) against EMI/EMC? |
Yes, we do have optimized “sniffer” probes for EMI and EMC work. These are near-field probes, both E-field and H-field that connect directly to instruments like oscilloscopes, VNAs, or spectrum analyzers. By positioning the probe around your board or system, you can quickly localize the physical source of emissions.
These probes are widely used for EMI troubleshooting, and we rely on them often. There are also several resources available, such as webinars, seminars, and application notes on our website and YouTube that show how to use these probes effectively. |
| Is it always true that ESR (package Zo) is optimal? For short connections, one would say that minimizing the ESR of the decap would work best. |
It’s not always true that the lowest ESR is the best choice. For instance, if you use a very low-ESR capacitor in a system where the board already has a resonance, it can actually make the problem worse by creating instability in the power supply.
In those cases, having a slightly higher ESR helps dampen the resonance by lowering the Q factor. Since multiple capacitors interact with each other along with their parasitics, it often becomes an optimization process. In fact, sometimes replacing a low-ESR capacitor with a higher-ESR one can eliminate a troublesome resonance altogether. |